A Broad Discussion of Speakers with Major Audio Luminaries

[audiofooled]

Oh, and to low frequency SPL: Consider barometric pressure changes. Yes, those are milliHz or microHz frequencies, but consider how big those change are. If you remember, approximately 194dB SPL is "1 atmosphere" amplitude. Driving from Estes Park Colorado to the Rocky Mtn National Park visitor center at the top is about a 3PSI change, if I recall correctly (it certainly lacks enough O2 for this human, let me tell you, wheeze wheeze). https://www.mide.com/air-pressure-at-altitude-calculator note, check your units FIRST!

3/14.7 = .2 .2 is about -6 db. If you didn't have your ear drum and eustachean tubes, you'd be dealing with a constant 188 dB stimulus at about 1/3600 Hz. Yeah, that eardrum is a useful thing, that.

How about a rate of change in pressure from high acceleration on the cone, or, for example, in low frequency percussion, mode 0 of the membrane that does not produce pitch or timbre qualities, but can be felt. Is there a point where the air itself becomes nonlinear?

 

[Neadych]

Bass is conventionally divided into three octaves. The lower range is 20-40 Hz. The middle range is 40-80 Hz. And the upper range is 80-160 Hz.
Does all of this account for 30% of our impressions of sound or not?

 

[MattHooper]

Brilliant description. This is something that confuses many listeners to believe that high distortion tube amps are more “dynamic” than solid state amps. And the FR modulation effect (by complex impedance) with seemingly more bass.

I’ve mentioned this before, but I have a feeling I have experienced something like that with my CJ tube amps when I had a more powerful Bryston 4B3 solid state amp for a few months to compare.

I had the sort of intuition that the Bryston was going to sound more powerful at least on some tracks and it didn’t quite work out that way.

For instance, one of my favourite pieces is the opening of the score for Conan the barbarian, which is a massively powerful orchestral piece, with an expanded Timpani section, expanded horn section, etc.

That had sounded really powerful with my CJ amps. Yeah when I switched in the Bryston amp (same volume) I was taking a back a bit because it didn’t seem to hit with the same punch and force. What it did instead was clean up and tighten the sound somewhat - when all the brass and strings and winds were playing together in the power region it could sound more muddy and congealed with my tube amp. Whereas with the Bryson amp, everything was tighter and better separated I could hear what was going on more between the sections. Also the powerful brass parts sounded a bit cleaner.

But the end result was that it sounded a bit less punchy and powerful.

In my work, I use compression fairly often and I recognized what seemed to be a sort of compression effect with the tube amps. It seemed to be the combination of less control in the upper/lower bass thickened that area and perhaps the less clean brass gave a bit of a compression effect, making the music feel both more punchy and louder.

I can’t say for sure that’s what was happening, but that was my strong impression, and it seems to fit with what is being said about perception/tube amp distortion.

BTW, where the Bryston did seem a bit more powerful was in a few different tracks.
For instance, some 80s highly syncopated synthesizer stuff, like Scritti Politti. In such songs the edge of the synths seemed a bit rounded and softened with my tube amps, and there’s lots of very sudden dynamic flashes of a guitar cord here or there, or a splashy synth hit, and those would seem to jump out with more solidity and dynamics in the Bryston.

 

[Floyd Toole]

Bass is conventionally divided into three octaves. The lower range is 20-40 Hz. The middle range is 40-80 Hz. And the upper range is 80-160 Hz.
Does all of this account for 30% of our impressions of sound or not?

The 30% statistic came from Sean Olive's research.
Olive, S.E. (2004a). “A multiple regression model for predicting loudspeaker preference using objective measurements: part 1 – listening test results”, 116th Convention, Audio Eng. Soc., Preprint 6113.

Olive, S.E. (2004b). “A multiple regression model for predicting loudspeaker preference using objective measurements: part 2 – development of the model”, 117th Convention, Audio Eng. Soc., Preprint 6190.

Or you can read summaries of it in my books.

The bass correlation was first noted in Part 2 of my 1986 paper where the -10 dB bass extension had a -0.5 correlation with sound quality ratings - all other factors varying! This means that bass extension, by itself is an important audible factor. -10 dB correlated better than -5 dB because of room gain.
Toole, F. E. (1986). “Loudspeaker measurements and their relationship to listener preferences”, J. Audio Eng. Soc., 34, pt.1, pp. 227-235, pt. 2, pp. 323-348.

What Sean Olive found in his first experiment using bookshelf loudspeakers with similar, but not identical, bass extension, was that bass accounted for 25% of the overall factor weighting, of which 6.27% was bass extension, and 18.64% was bass quality (frequency response). In his second experiment which used 70 loudspeakers of all sizes the bass extension alone accounted for 30.5% of the overall factor weighting. The variations in bass extension were noted and appreciated. Smoothness very likely was a factor, but in these comparisons was less noticed.

The same phenomenon was noticed in an elaborate comparison of rooms and loudspeakers:

Schuck, P., Olive, S., Toole, F., Sally, S., Bonneville, M., Momtahan, K. and Verrault, E. (1993). “Perception of Perceived Sound in Rooms: Some Results of the Athena
Project”, 12th International Conference, Audio Eng. Soc., Paper 12-006. and

Olive, S.E., Schuck, P.L., Sally, S.L. and Bonneville, M. (1995). “The Variability of Loudspeaker Sound Quality Among Four Domestic Sized Rooms”, 99th Convention, Audio Eng. Soc., Preprint 4092.

In these experiments, when binaural comparisons of loudspeaker and rooms combined were done (i.e. listeners could not perceptually separate the loudspeaker from the room), differences between the loudspeakers were not significant and the room dominated. The most perferred loudspeaker/room combinations were those with most bass, even when the bass was obviously boomy. Listeners were less discriminating in their assessments of the loudspeakers when bass was a significant perceptual variable.

People like bass, even bad bass is preferred to insufficient bass.

So the answer to your question is that the lower the cutoff frequency the better, and above that smoothness matters. Small room resonances are the problem in the real world. and the lower the cutoff frequency the more of them are energized. OK?

 

[Floyd Toole]

I’ve mentioned this before, but I have a feeling I have experienced something like that with my CJ tube amps when I had a more powerful Bryston 4B3 solid state amp for a few months to compare.

I had the sort of intuition that the Bryston was going to sound more powerful at least on some tracks and it didn’t quite work out that way.

For instance, one of my favourite pieces is the opening of the score for Conan the barbarian, which is a massively powerful orchestral piece, with an expanded Timpani section, expanded horn section, etc.

That had sounded really powerful with my CJ amps. Yeah when I switched in the Bryston amp (same volume) I was taking a back a bit because it didn’t seem to hit with the same punch and force. What it did instead was clean up and tighten the sound somewhat - when all the brass and strings and winds were playing together in the power region it could sound more muddy and congealed with my tube amp. Whereas with the Bryson amp, everything was tighter and better separated I could hear what was going on more between the sections. Also the powerful brass parts sounded a bit cleaner.

But the end result was that it sounded a bit less punchy and powerful.

In my work, I use compression fairly often and I recognized what seemed to be a sort of compression effect with the tube amps. It seemed to be the combination of less control in the upper/lower bass thickened that area and perhaps the less clean brass gave a bit of a compression effect, making the music feel both more punchy and louder.

I can’t say for sure that’s what was happening, but that was my strong impression, and it seems to fit with what is being said about perception/tube amp distortion.

BTW, where the Bryston did seem a bit more powerful was in a few different tracks.
For instance, some 80s highly syncopated synthesizer stuff, like Scritti Politti. In such songs the edge of the synths seemed a bit rounded and softened with my tube amps, and there’s lots of very sudden dynamic flashes of a guitar cord here or there, or a splashy synth hit, and those would seem to jump out with more solidity and dynamics in the Bryston.

Did you take into account the fact that most tube amps have sufficiently high output impedances that they can audibly change the sound of loudspeakers? I don't know about your CJs. Do you?

 

[j_j]

How about a rate of change in pressure from high acceleration on the cone, or, for example, in low frequency percussion, mode 0 of the membrane that does not produce pitch or timbre qualities, but can be felt. Is there a point where the air itself becomes nonlinear?

At 140dB SPL very much so. Below that there is some dispute, and some need to specify the dP/dt involved, I believe. Remember that we are talking about directly at the loudspeaker, not at the listener, so the levels can be much higher there, especially if horns and horn throats are involved, or if small volume enclosures with bass in them.

 

[olieb]

In his second experiment which used 70 loudspeakers of all sizes the bass extension alone accounted for 30.5% of the overall factor weighting.

In these experiments, when binaural comparisons of loudspeaker and rooms combined were done (i.e. listeners could not perceptually separate the loudspeaker from the room), differences between the loudspeakers were not significant and the room dominated. The most perferred loudspeaker/room combinations were those with most bass, even when the bass was obviously boomy.

So if one were to follow preference statistics, it seems like we want
A. deep bass extension
B. a lot of bass (even if it gets boomy)
C. many channels (as this will blur differences in speaker "quality" to a high degree)
EDIT:
D. Flat on axis (though this is of somewhat limited importance because of C.)

Is this really the solution? Or is there a problem with preference statistics?

And how can the room dominate the speaker when room reflections are not doing that much?

I attach my version, and you will see that even the first room reflections are around the threshold of detection for image shift. So, don't get concerned about "image shift and focus" caused by even later room reflections.

Is the (strong) room influence all about unevenness in FR from small room problems?

 

[respice finem]

Statistics are what they are - statistics. I for one don't subscribe to B. Boomy bass?
I like slightly lean bass, not masking the mids, better.

The nice side of this hobby (for me): no one must follow the mainstream, if its "teachings" don't suit them.

 

[MattHooper]

Did you take into account the fact that most tube amps have sufficiently high output impedances that they can audibly change the sound of loudspeakers? I don't know about your CJs. Do you?

Yes.

My CJs we’re measured by Stereophile and they have sufficiently high impedance to interact audibly with some loudspeakers, and I have several loudspeakers some of which would challenge those amplifiers.
That’s on top of some of the measured added distortion characteristics.
So I include that as a possible factor, and I believe that is where some of the audible differences were coming in - like the thickened bulge in the upper bass there or thereabouts.

The amplifier comparisons were not done blind, so there’s that as well.

(On the other hand: I’ve also done blind testing of my CJ tube preamplifier against my Benchmark solid state preamplifier. Level matched with voltmeter at the speaker terminals, randomized switching - in two sets of 15 trials, was able to reliably identify each - 14/15 correct, 15/15 correct).

 

[Floyd Toole]

So if one were to follow preference statistics, it seems like we want
A. deep bass extension
B. a lot of bass (even if it gets boomy)
C. many channels (as this will blur differences in speaker "quality" to a high degree)
EDIT:
D. Flat on axis (though this is of somewhat limited importance because of C.)

Is this really the solution? Or is there a problem with preference statistics?

And how can the room dominate the speaker when room reflections are not doing that much?

Is the (strong) room influence all about unevenness in FR from small room problems?

Statisitical data are indicators of possibly reliable trends. The human behaviour we are attempting to decipher is multidimensional and the dimensions are interactive. However, over time one sees that same trends reappearing often enough that one can start placing bets. Along the way, it has been observed numerous times that humans adapt, become familiar with and normalize, sometimes substantial flaws in loudspeakers and rooms. That is why we do double-blind, equal-loudness multiple-comparison tests. These are costly and time consuming, which is why they are rarely done. They provided the raw data for the identification of meaningful measurements. Now a timbrally neutral, resonance-free, loudspeaker is more reliably identified from inspecting a spinorama than by listening under the circumstances available to most people.

Based on my observations thus far, my bets would be on:
A. As much bass extension as makes sense. By that I mean, if you are into organ recitals, 16 Hz might occasionally be called upon. Explosions in movies seem to benefit from very low bass extension. Modern pop music with synthesizers can also plumb the very low frequencies - I have heard/felt and measured 20 Hz. Did it add to the music? Not really, and I wonder if the creators actually heard it. Traditional Rock and roll, not so much - punchy 80 Hz is appreciated. And so on. At Harman I recall the Infinity Prelude MTS with its dedicated powered subwoofer reached about 22 Hz at -10dB (see Figure 5.6 in the 3rd edition for the spinorama). For the time (ca. 2000) it was an exceptional loudspeaker and it was occasionally statistically tied in double-blind multiple comparison tests, but never bettered. It is an example of the state of the art in linear performance at the time.
B. Enough bass quantity to be gratifying for the intended audience. I know there are "bassaholics" out there for whom there may be no limits. I like bass, and I like it better when it is non-resonant. There are solutions for small room resonances and seat-to-seat variations, and Chapter 14 in the 4th edition gives this expanded treatment thanks to the input of Todd Welti. Would I tolerate a bit of "boom"? No, because have heard better and know how to fix it. Many listeners are simply thrilled to hear any low bass.
C. How many channels? More are better, but not because the recorded large room resonances reduce our ability to detect resonances in loudspeakers. Besides all stereo and multichannel programs have hard-panned, monophonic, components where the resonances can be heard. These days, fundamentally neutral, resonance-free, loudspeakers exist in large numbers and attractive prices. Spectral tilts and tone-control variations are still necessary to address program variations if one is truly fussy.
D. The direct sound is still the dominant factor in sound localization, including phantom images, and in sound quality above the transition frequency. It needs to be flattish and smooth, indicators of the absence of resonances. Off-axis spatially averaged data confirm the existence of resonances as distinct from acoustical interference.

"And how can the room dominate the speaker when room reflections are not doing that much? Is the (strong) room influence all about unevenness in FR from small room problems?" I'd bet on it.

 

[pma]

@MattHooper , you have this?

Conrad-Johnson LP125M monoblock power amplifier Measurements | Stereophile.com

Sidebar 3: Measurements To take these measurements, I used Stereophile's loan sample of the top-of-the-line Audio Precision SYS2722 system (see the January 2008 "As We See It" and www.ap.com). The sample shipped to me for measurement had its output transformer set for an 8 ohm load. Before...

www.stereophile.com

Or MV55? Anyway, the modulated FR shown is a good candidate for colorations. There is a way to make speaker complex impedance closer to resistive one, however at the expense of lower resulting impedance.

 

[MattHooper]

@MattHooper , you have this?

Conrad-Johnson LP125M monoblock power amplifier Measurements | Stereophile.com

Sidebar 3: Measurements To take these measurements, I used Stereophile's loan sample of the top-of-the-line Audio Precision SYS2722 system (see the January 2008 "As We See It" and www.ap.com). The sample shipped to me for measurement had its output transformer set for an 8 ohm load. Before...

www.stereophile.com

Or MV55? Anyway, the modulated FR shown is a good candidate for colorations. There is a way to make speaker complex impedance closer to resistive one, however at the expense of lower resulting impedance.

My current amplifiers, which I’ve had for about 25 years now, are the Conrad Johnson premier 12 tube monos. MEASUREMENTS HERE

(as it happens, previous to the premier 12s, I did own the less powerful MV55 CJ stereo amp, which seemed to run out of steam once I started taxing them with more full range and complex loudspeakers) .

 

[j_j]

@MattHooper , you have this?

Conrad-Johnson LP125M monoblock power amplifier Measurements | Stereophile.com

Sidebar 3: Measurements To take these measurements, I used Stereophile's loan sample of the top-of-the-line Audio Precision SYS2722 system (see the January 2008 "As We See It" and www.ap.com). The sample shipped to me for measurement had its output transformer set for an 8 ohm load. Before...

www.stereophile.com

Or MV55? Anyway, the modulated FR shown is a good candidate for colorations. There is a way to make speaker complex impedance closer to resistive one, however at the expense of lower resulting impedance.

Without naming names, kof kof, some nominal 8-ohm speakers that drop to 2 ohms around a crossover point really **** from the amplifier's POV.

 

[MarkS]

My CJs we’re measured by Stereophile and they have sufficiently high impedance to interact audibly with some loudspeakers

Reminds me of the Carver challenge:

Carver Challenge | BobCarverCorp.com

www.bobcarvercorp.com

I'm happy to just fiddle around with PEQ on my Lyngdorf amp ...

As for added nonlinear distortion, I Just Say No!

 

[MattHooper]

As for added nonlinear distortion, I Just Say No!

Har. You’ll have to take my nonlinear distortion from my cold, dead hands.

 

[Arindal]

Modern pop music with synthesizers can also plumb the very low frequencies - I have heard/felt and measured 20 Hz. Did it add to the music? Not really, and I wonder if the creators actually heard it.

I would argue, they did hear it in most cases, otherwise they would not have added inaudible frequencies which eat up a lot of energy. We should take into account that lots of popular music, particularly EDM and hiphop, are mixed and mastered not only (or not primarily) for home stereo systems, but for club P.A., headphones and IEM. These tend to have no problems with very low bass, and particularly DJs were demanding proper reproduction from the mid-1990s on. Since the beginning of the ´Beats by Dr. Dre´ era, i.e. since 2008, consumers seemingly pay some attention to that as well.

Nowadays, very low bass extension is pretty common, and obviously plays an important role for buyers. I recently came across a very compact portable bluetooth speaker, which had an estimated effective volume of 0.9l (!) per bass driver, but reached an astonishingly low 32Hz (-3dB).

 

[Floyd Toole]

I would argue, they did hear it in most cases, otherwise they would not have added inaudible frequencies which eat up a lot of energy. We should take into account that lots of popular music, particularly EDM and hiphop, are mixed and mastered not only (or not primarily) for home stereo systems, but for club P.A., headphones and IEM. These tend to have no problems with very low bass, and particularly DJs were demanding proper reproduction from the mid-1990s on. Since the beginning of the ´Beats by Dr. Dre´ era, i.e. since 2008, consumers seemingly pay some attention to that as well.

Nowadays, very low bass extension is pretty common, and obviously plays an important role for buyers. I recently came across a very compact portable bluetooth speaker, which had an estimated effective volume of 0.9l (!) per bass driver, but reached an astonishingly low 32Hz (-3dB).

Good points. But forcing systems to reproduce energetic sub-musical frequencies not only uses energy, but very often generates non-linear distortion. It requires a good loudspeaker system, even decent headphones, not to distort - boom cars included. Of course professional users prepare for this. Personally, I enjoy extended bass, but I have had good gear. The only downside is that one occasionally hears very low-frequency sounds that shouldn't be there.

Yes, I too have experienced unexpectedly good bass from small portable speakers. That is the result of integrated amplifiers, digital EQ and protective algorithms that progressively high-pass the bass as the volume is increased. For background music they are definitely impressive.

The same technology is in powered pro monitors and consumer loudspeakers and the result is even more neutral sound than passive systems normally deliver. It is a pity that consumers are reluctant to use active "HiFi" systems, but that may change.

 

[DanielT]

I recently came across a very compact portable bluetooth speaker, which had an estimated effective volume of 0.9l (!) per bass driver, but reached an astonishingly low 32Hz (-3dB).

And the maximum SPL of that speaker is?

It would be interesting to see a distortion - SPL graph on those speakers.

 

[Arindal]

And the maximum SPL of that speaker is?

In the ballpark of 93dB, but that naturally comes with a gradually-deployed high-pass filter, increasing the lower extension to 55Hz at full throttle. It was not my favorite bass in terms of quality, with a hint of boominess, nevertheless impressive.

I too have experienced unexpectedly good bass from small portable speakers. That is the result of integrated amplifiers, digital EQ and protective algorithms that progressively high-pass the bass as the volume is increased. For background music they are definitely impressive.

It is not only highpass filters of dynamically increasing cutoff frequency, but also some narrow band-selective limiters (reducing the bands which cause highest excursion), soft-knee compressors or alike to make the bass ´fatter´ over time before the system runs out of power of excursion capacity. Most of such systems seemingly rely on passive radiators, not closed or vented box designs.

The same technology is in powered pro monitors and consumer loudspeakers and the result is even more neutral sound than passive systems normally deliver. It is a pity that consumers are reluctant to use active "HiFi" systems, but that may change.

Interestingly, I see more and more ´non-traditional-hifi´ consumers tending to active speakers, oftentimes with built-in network capabilities. Interestingly, the majority of such popular products (like B&O, Bose, Sonos, Blusound, KEF, Technics, Marantz, Devialet etc.) relies on either closed-box designs with a whole lot of power, or vented boxes. Both is coming with certain limitations in very compact designs, and seemingly use dynamic DSP only to a certain degree, so I never really met an impressive performance-to-size ratio.

As all of these components used in portable speakers, exist in the OEM sphere and are seemingly not as expensive (otherwise they would not end up in $200 or $300 products), I wonder why they are not more frequently used in more expensive yet compact active stereo speakers for home use.

 

[DanielT]

In the ballpark of 93dB, but that naturally comes with a gradually-deployed high-pass filter, increasing the lower extension to 55Hz at full throttle. It was not my favorite bass in terms of quality, with a hint of boominess, nevertheless impressive.

Interesting.

My niece bought a compact portable bluetooth speaker a while ago. She might have thought about this one you're talking about if she knew about it. Or if I had suggested it to her.Now I got curious,what manufacturer/model is it?